Locke virtually recognizes the extreme importance of the doctrine of sensation, and his second book might almost be entitled "Concerning the Nature and Products of Sensation." On the other hand, one of the most characteristic and valuable portions of the reply of Leibniz is in his development of a theory of sensation which is thoroughly new, except as we seek for its germs in its thoughts of Plato and Aristotle. According to Locke, knowledge originates from two sources,--sensation and reflection. Sensations are "the impressions made on our senses by outward objects that are extrinsic to the mind." When the mind "comes to reflect on its own operations about the ideas got by sensation, and thereby stores itself with a new set of ideas," it gets ideas of reflection. If we leave out of account for the present the ideas of reflection, we find that the ideas which come through sensation have two main characteristics. First, in having sensations, the mind is passive; its part is purely receptive. The objects impress themselves upon the mind, they obtrude into consciousness, whether the mind will or not. There is a purely external relation existing between sensation and the understanding. The ideas are offered to the mind, and the understanding cannot refuse to have them, cannot change them, blot them out, nor create them, any more than a mirror can refuse, alter, or obliterate the images which objects produce in it. Sensation, in short, is a purely passive having of ideas. Secondly, every sensation is simple. Locke would say of sensations what Hume said of all ideas,--every distinct sensation is a separate existence. Every sensation is "uncompounded, containing nothing but one uniform appearance, not being distinguishable into different ideas." Knowledge is henceforth a process of compounding, of repeating, comparing, and uniting sensation. Man's understanding "reaches no further than to compound and divide the materials that are made to his hand." It hardly need be said that Locke has great difficulty in keeping up this thoroughly atomic theory of mind. It is a theory which makes all relations external; they are, as Locke afterwards says, "superinduced" upon the facts. It makes it impossible to account for any appearance of unity and connection among ideas, and Locke quietly, and without any consciousness of the contradiction involved, introduces certain inherent relations into the structure of the ideas when he comes to his constructive work. "Existence and unity are two ideas," he says, "that are suggested to the understanding by every object without, and every idea within."